The invention is directed to novel materials and their use for the electrocatalytic production and oxidation of H₂. More particularly, it concerns materials which can be used for the production of electrodes in the field of electronics, and notably electrodes for fuel cells, electrolysers and photoelectrocatalytical (PEC) devices.

Molecular hydrogen is widely considered as a convenient energy vector. Its combustion in a fuel cell generates electricity with high yield and without any pollutant exhaust, water being the sole reaction product. However, although hydrogen is one of the most abundant elements on earth, molecular hydrogen exists only as traces in the atmosphere and has to be produced through processes that require some energy input. Thus the economically viable production of H₂ from renewable sources is a major concern to the scientific community. Electrolysis (reduction of proton into H₂ by mean of electrical energy) or photoelectrolysis (reduction of proton into H₂ by mean of light energy sometimes with additional electrical energy) is one way to produce H₂ from water (or a convenient proton source). Reduction of protons is apparently a very simple reaction. Unfortunately, it progresses slowly on most electrodes except on noble metals such as platinum. Hence hydrogen evolution is generally not observed at potentials near equilibrium (-400 mV vs. SHE at pH 7 in water) but requires the application of an overpotential also called activation potential. The same occurs for hydrogen oxidation. This kinetic limitation thus significantly reduces the energetic yield during a complete formation/uptake cycle of hydrogen and is therefore economically limiting for most industrial applications.

Fuel cells are electrochemical devices that convert the energy of a fuel directly into electrical and thermal energy. Generally, a fuel cell consists of an anode and a cathode, separated by an electrolyte through which they are electrically connected. A fuel, usually hydrogen, is fed to the anode where it is oxidised with the help of an electrocatalyst. At the cathode, the reduction of an oxidant such as oxygen, and usually air, takes place. The electrochemical reactions at the electrodes produce an electrical current and therefore produce electrical energy. Usually, thermal energy is also produced in parallel and can be used to provide additional electricity or for other purposes.

Currently the most common electrochemical reaction which is performed in a fuel cell is the reaction between hydrogen and oxygen to produce water. Molecular hydrogen may be fed to the anode where it is oxidised, the electrons produced passing through an external circuit to the cathode where the oxidant is reduced. Ions flowing through the intermediate electrolyte maintain charge neutrality. Fuel cells can be adapted to use other fuels such as methanol, hydrazine or natural gas.

Water electrolysers are made as fuel cells but are operated the reverse, with the application of electrical power between the electrodes and supply with water. Hydrogen is generated at the cathode and oxygen is generated at the anode.

Various difficulties have prevented the commercial development of fuel cell and water electrolyser technologies. The first one is cost and, in particular, the cost of the electrocatalysts employed at both the cathode and the anode to facilitate the electrochemical interconversion of 2H⁺ + 2e^- into H₂ in the most versatile technology based on proton exchange membrane such as Nafion ®. The most commonly used electrocatalyst is platinum. Platinum is a very efficient catalyst and enables high currents to be produced in the fuel cell. However, its cost is high, and platinum is of limited availability. Therefore the metal catalyst is a significant contributor to the expense of the fuel cell. A further difficulty is that platinum is available in limited quantities on earth and world supplies cannot be expected to last more than a few decades if the use of fuel cells was to be generalized, in cars notably (Gordon et al. Proc. Natl Acas Sci USA, 2006, 1209-1214). The active layers for either hydrogen electro-evolution or electro-oxidation contain nanoparticles of Pt or other noble metals coated onto a carbon material: Pt nanoparticle catalysts stem from mid 20^th century research on chemical catalysis in the gas phase and have been largely optimized up to now. While this solution is today the sole which is economically viable, nanoparticles (usually with a diameter of 5nm and more) present no more than 10% of their atoms on their surface, while 90% of metal weight is unnecessarily immobilised, resulting in a cost/efficiency unfavourable balance.

Another difficulty with platinum comes from the fact that it is irreversibly inactivated in the presence of carbon monoxide. Many sources of hydrogen gas contain carbon monoxide impurities. The use of platinum catalysts therefore requires hydrogen fuel of high purity, with extremely low carbon monoxide levels. This adds to the cost of operating the fuel cell.

As a consequence, the existence of a global hydrogen economy appears to be dependent upon the development of new base-metal catalysts for hydrogen production and uptake.

Alternatives to platinum catalysts have been the object of numerous studies: one possibility is the use of hydrogenase enzymes at the anode. Electrodes based on the combination of hydrogenase enzymes with a carbon material have been disclosed in WO 2003/019705 ; US 2007/0248845 ; EP1939961 ; WO2004/114494, and more extensively in J.A. Cracknell et al., Chem. Rev. 2008, 108, 2439-2461. However, hydrogenases have been found to be highly sensitive to the presence of oxygen, and become inactive over a period of time when used in a standard fuel cell operating with oxygen (or an oxygen containing material such as air) as the oxidant. Moreover, they are very difficult to produce in a catalytically active form in significant amounts. A representative preparation requires two weeks for a few milligrams of enzymes corresponding to a few amount of active molecule since the molecular weight of the catalyst is about 55 kg.mol^-1.

US2006/0093885 discloses functionalized carbon materials which can be used in electrochemical assemblies for electrochemical cells or fuel cells. But the catalyst is based on noble metal particles dispersed in a polymer.

WO2006063992 and US2006/0058500 disclose metal particles formed from a polymeric resin, a transition metal and a reducing agent and their use to make electrodes for fuel cells.

WO2006074829 discloses metal-organic complexes which are used as a coating on an anion conducting membrane, further reduced to form metal particle and fuel cells made from such membranes.

In these three last documents, metal ions are reduced to form nanoparticles and catalytic activity relies on the use of these nanoparticles.

The immobilization of metal-organic complexes on the surface of electrodes has also been achieved with the same aim: Kellet, R. M. and Spiro, T. G., (Inorg. Chem. 1985, 24, 2378) reported a series of cobalt porphyrins with high activity for H₂ production, associated with low overvoltage in neutral aqueous solution. However, these compounds proved difficult to handle when covalently grafted to an electrode via an amide link with surface carboxylic acid groups: film instability or disrupting processes at either film-electrode or film-electrolyte interfaces were postulated. In another example, incorporating positively charged cobalt porphyrin complexes into a Nafion ® membrane coated on a glassy carbon electrode resulted in low electroactivity, reflecting the poor electron-transfer characteristics of Nafion ® films. T. Abe et al. (Polym. Adv. Technol. 1998, 9, 559) reported that cobalt tetraphenylporphyrin incorporated in a Nafion ® membrane coated on a bare pyrolytic graphite electrode can reduce protons only with a larger overvoltage (-0.7 V vs Ag/AgCl) in a pH 1 aqueous solution together with a considerably lower 70 h^-1 turnover frequency value. Better turnover frequency (2.10⁵ h^-1) was observed but still with a large overvoltage at a potential of -0.90 vs Ag/AgCl and pH =1 for a cobalt phthalocyanine incorporated in a poly(4-vinylpyridine-co-styrene) film coated on a graphite electrode. In that case again the catalytic proton reduction was limited by the electron transfer within the matrix (Zhao et al. J. Mol. Catal. A 1999, 145, 245). Electropolymerization of [Cp*Rh(L)Cl](BF₄) (L = bis-4,4'-bispyrrol-1-ylmethyl)methoxycarbonyl]-2,2'-bipyridyl) leads to a stable film capable of proton electroreduction. Quantitative current efficiency corresponding to 353 turnovers was observed during a 14 hours electrolysis experiment at pH 1 using a carbon-felt electrode coated with the electropolymerized rhodium complex. Here again, turnover frequency is low and overvoltage dominates (Cosnier et al. J. Chem.Soc. Chem. Comm. 1989, 1259). More recently, the immobilization of diiron complexes (also known to be molecular electrocatalysts for hydrogen evolution) on carbon materials, either via the electropolymerisation of diazonium salts on glassy carbon (V.Vijaikanth et al. Electrochemistry communication, 2005, 7, 427-430) or via polypyrrole coating (S.K.Ibrahim et al. Chem Commun 2007, 1535-1537) has been described. However, although diiron complexes in solution show promising electrocatalytic activity, there is little or no activity retained after grafting.

Today, there are no other solutions to efficiently replace platinum as electrocatalyst at the anode of a hydrogen fuel cell.

It has been an aim of this invention to provide a solution to the problems of the prior art, and notably: to provide a material which can be used as electrocatalyst in fuel cells and electrolysers, which is easy to prepare, based on cheap materials, and permits the economically viable production and oxidation of H₂, which implies a catalyst with a low activation potential, and a resistance to degradation under working conditions.

The solution to this problem is a new material comprising a classical electrode solid support, on which a coordination complex is grafted through a linker arm.

It has been found that, in the material of the invention, metal-organic complexes retained their electrocatalytic activity once grafted on the solid support thanks to the use of selected linkers. Furthermore, the immobilization, when carried out via this procedure, does not result in an increased overvoltage as compared to that observed for the catalyst in solution. Materials of the invention differ from prior art materials showing little or no catalytic activity notably by the nature of the linker arm.

The material of the invention is easy to prepare, can be integrated in a fuel cell or in an electrolyser as an electrode. It works as molecular electrocatalyst for hydrogen evolution or oxidation, and is characterized by an improved stability (characterized by the total number of catalytic cycles achieved). In addition, the catalytic layer can be made transparent so that it can be used as cathode in a photoelectrocatalytical device.

A first object of the invention is a material comprising a solid support of a conductive or semi-conductive material which is functionalized on its surface by linker arms, said linker arms comprising at least two extremities, wherein the first extremity is bonded to the solid support either by a covalent link or via π-stacking interaction and the second extremity is linked in a covalent manner to a metal-organic complex (C*). Such a material is schematically illustrated on schema A here-under.

The solid support preferably has the shape of a plaque, but this shape can be different. Notably, the solid support can have the shape of a rod, a cylinder or wire, according to the device in which it is used.

The solid support advantageously is a conductive or semi-conductive material with a high specific surface. It may be nano-structured or not. This conductive or semi-conductive material with a high specific surface may be deposited on another support of a conductive material, in order to form an electrode with a high specific surface. This other conductive material may be made of any conducting material, for example indium tin oxide (ITO), stainless steel, iron, copper, nickel, cobalt, aluminium (in particular when it is freshly brushed), gold, doped-diamond, titanium, brass or carbon, e.g. graphite. Preferably it is made of a material selected from ITO and graphite.

Materials for the preparation of the solid support can be selected from: a metallic material, a carbon material, a semi-conductor or conductor metal oxide, nitride or chalcogenide.

When the solid support is a metallic material, it can be selected from: silicon, brass, stainless steel, iron, copper, nickel, cobalt, aluminium (in particular when it is freshly brushed), silver, gold or titanium.

Functionalisation on a metallic solid support is performed by a reaction with an aryl radical from the linker molecule leading to covalent grafting.

When the solid support is a carbon material, it can be selected from:

• a curved carbon nanostructure like carbon black, single or multi-walled carbon nanotubes (CNT), fullerenic nanoparticles,
• graphite,
• glassy carbon (expanded or not, or a foam),
• graphene,
• doped diamond,

A curved carbon nanostructure includes, but is not limited to, a carbon nanotube (CNT), a fullerenic nanoparticle and carbon black.

Methods for the preparation of CNT include laser vaporization of graphite (Thess et al, Science 273, 1996, 483), arc discharge (Journet et al, Nature 388, 1997, 756) and the HiPCo (high pressure carbon monoxide) process (Nikolaev et al, Chem. Phys. Lett. 313, 1999, 91-97). Other methods for producing CNTs include chemical vapor deposition (Kong et al, Chem. Phys. Lett. 292, 1998, 567-574; Cassell et al, J. Phys. Chem. 103, 1999, 6484-6492); and catalytic processes both in solution and on solid substrates (Yan Li et al, Chem. Mater. 13(3); 2001, 1008-1014 ; A. Cassell et al, J. Am. Chem. Soc. 121, 1999, 7975-7976).

Methods for preparing fullerenic nanoparticles are described in US-5,985,232 and US 2004/057,896. Fullerenic nanoparticles are available commercially from suppliers such as Nano-C Corporation, Westwood Mass.

Carbon black is a powdered form of highly dispersed, amorphous elemental carbon. It is a finely divided, colloidal material in the form of spheres and their fused aggregates. Different types of carbon black can be distinguished by the size distribution of the primary particles, and the degree of their aggregation and agglomeration.

Carbon black is made by a method selected from controlled vapor-phase pyrolysis and/or thermal cracking of hydrocarbon mixtures such as heavy petroleum distillates and residual oils, coal-tar products, natural gas and acetylene. Acetylene black is the type of carbon black obtained by the burning of acetylene. Channel black is made by impinging gas flames against steel plates or channel irons, from which the deposit is scraped at intervals. Furnace black is the name of carbon black made in a refractory-lined furnace. Lamp black is made by burning heavy oils or other carbonaceous materials in closed systems equipped with settling chambers for collecting the solids. Thermal black is produced by passing natural gas through a heated brick checkerwork where it thermally cracks to form a relatively coarse carbon black. Carbon black is available commercially from many suppliers such as Cabot Corp.

A fullerene is a spherical allotrope of carbon. It is a molecule composed entirely of an even number of carbon atoms in the sp²-hybridized state and takes the form of a closed cage. The most abundant species is the C60 molecule, and the second most abundant species of the fullerene family is C70. Fullerenes include not only single-walled but also multi-walled cages consisting of stacked or parallel layers. The preparation of fullerenes has been disclosed notably in WO 92/04279, US-5,316,636, US- 5,300,203, and Howard et al., Nature 352, 1991, 139-141. Fullerenes may be obtained commercially from suppliers such as Carbon Nanotechnologies Incorporated, MER Corporation, Nano-C Corporation, TDA Research Inc., Fullerene International Corp., and Luna Innovations.

In order to be functionalized, carbon materials used in this invention include carbon materials having unsaturations (graphite, graphene, carbon nanotubes, carbon black, glassy carbon). They are functionalized in a known manner:

• by addition chemistry performed on one or more C=C double bonds at the surface. Aryl radicals, generated by reduction of diazonium salts from linker arm precursors, interact with carbon material in this way.
• via π-stacking interaction of graphene layers of the carbon materials with aromatic moieties of the linker arms such as phenyl, biphenyl, anthracene, pyrene, perylene, and other polyaromatic groups.

When the solid support is a semi-conductor or conductor metal oxide, nitride or chalcogenide, it can be selected from: TiO₂, NiO, ZnO, ZrO₂ ITO, SnO₂, WO₃, Fe₂O₃, Ta₂O₅, Ta₃N₅, TaON, N-doped TiO₂, ZnS, ZnSe, CdS, CdSe, CdTe, ZnTe and composites of these materials, also possibly doped with other elements. Favourite metal oxide is selected from: NiO, TiO₂, ZnO.

Functionalization of the solid support in that case is performed by a reaction/interaction such as: a covalent link between an aryl radical of the linker molecule with the metal oxide materials.

Favourite material for the solid support is selected from: multi-wall carbon nanotubes (MWCNT), and metal oxide. Even more preferably, the solid support is selected from multi-wall carbon nanotubes (MWCNT).

The linker arms comprise at least two extremities: the first extremity is bonded to the solid support either by a covalent link or via π-stacking interaction and the second extremity is linked in a covalent manner to a metal-organic complex.

The linker should advantageously be long enough to let the molecular compound adopt the conformation required for catalysis. And preferably, the linker should be not too long, in order to avoid its forming an insulating layer at the surface of the conductive or semi-conductive material and hinder the electron transfer from the electrode to the catalyst.

Said linker arm comprises a first extremity which is bonded to the solid support in a manner that insures good electronic transfer with low resistivity and the second extremity is linked in a covalent manner to a metal-organic complex. Good electronic transfer between the solid support and the metal-organic complex can be assessed via the measurement of the electric resistivity of the layer. Low resistivity is achieved thanks to covalent linkage or π-stacking interactions between the linker and the solid support.

The linker can be selected from hydrocarbon molecules, said hydrocarbon molecule comprising at least one covalent link with the solid support or one group interacting through π-stacking with the solid support, and comprising also at least one covalent link with the metal-organic complex.

This hydrocarbon molecule is selected from those responding to the following formulae: and to oligomers resulting from the oligomerization of: wherein:

• Ar represents a C₆-C₃₀ aromatic residue, possibly comprising one or more substituants selected from: C₁-C₆ alkyl, -OH, -NH₂, -COOH, -F, -Cl, -Br, -I, -NO₂, -CONR''₂, -COOR", -SO₃^-, -SR", -OR", -NR''₂;
• R and R', identical or different, represent a group selected from : H, C₁-C₂₀ alkyl, C₁-C₂₀ alkenyl, C₁-C₂₀ alkynyl, C₆-C₃₀ aryl, C₆-C₃₀ aralkyl, possibly comprising one or more substituants selected from: -OH, -NH₂, -COOH, -CONH₂, triazole, -SH, -N₃, and possibly interrupted by one or more bridges selected from :-CONH-, -CO-O-, -CO-O-CO-, -CO-NH-CO-,-CO-S-, -CS-O-, -CS-S-;
• R" represents a group selected from: H, C₁-C₂₀ alkyl, C₁-C₂₀ alkenyl, C₁-C₂₀ alkynyl, C₆-C₃₀ aryl, C₆-C₃₀ aralkyl;
• X represents a group selected from: C₁-C₂₀ alkyl, C₁-C₂₀ alkenyl, C₁-C₂₀ alkynyl, C₆-C₃₀ aryl, C₆-C₃₀ aralkyl, possibly comprising one or more substituants selected from: -OH, -NH₂, -COOH, -CONH₂, triazole, -SH, -N₃, and possibly interrupted by one or more bridges selected from :-CONH-, -CO-O-, -CO-O-CO-, -CO-NH-CO-,-CO-S-, -CS-O-, -CS-S-;
• x is an integer, 1 ≤ x ≤5, x is the number of functional groups Y which are grafted on the X group;
• Y is a functional group selected from: a simple covalent link, -O-, - NH-,-S-, -COO-, -CONH_-, -CO-S-, -CS-O-, -CS-S-.

Favourite linker arms are those resulting from the oligomerization of

When the linker is the interaction between the linker and the solid support is via π-stacking.

When the linker is an oligomer resulting from the oligomerization of: the linker is covalently linked through the Ar group, to the solid support.

When the linker is the interaction between the linker and the solid support is a covalent bond resulting from a reaction between one functional group at the surface of the solid support and the linker molecule.

In formulae (IA), (IB), and (IC), the favourite variants are the following:

• Ar advantageously represents a C₆-C₃₀ aromatic residue. Ar can be selected from the following list: phenyl, biphenyl, pyrenyl, anthracenyl, phenanthrenyl, perylenyl, naphtacenyl.

Preferably R and R' represent a group selected from: H, C₁-C₂₀ alkyl, C₁-C₂₀ alkenyl, C₁-C₂₀ alkynyl, C₆-C₃₀ aryl, C₆-C₃₀ aralkyl, and even more preferably selected from: H, C₁-C₁₂ alkyl. Advantageously R=R'=H.

Preferably X represents a group selected from: C₁-C₂₀ alkyl, C₁-C₂₀ alkenyl, C₁-C₂₀ alkynyl, C₆-C₃₀ aryl, C₆-C₃₀ aralkyl. Even more preferably X represents a group selected from: C₁-C₁₂ alkyl, C₁-C₁₂ alkenyl, C₁-C₁₂ alkynyl, C₆-C₂₀ aryl, C₆-C₂₀ aralkyl.

Preferably 1≤ x ≤3, even more preferably x=1.

Alkyl, alkenyl and alkynyl groups can be linear, branched or cyclic.

Said linker molecule can take different shapes: it can be simply a linear chain linking the solid support and the metal-organic complex, but it can also be a branched molecule comprising several arms, one or several arms being linked to the solid support and one or several arms being linked to the metal-organic complex, while other arms can remain free. This last case is especially illustrated in the experimental part wherein the linker is an oligomer resulting from the oligomerization of:

The linker is not exclusively a definite compound. During fabrication, some oligomerization of the linker can occur but this process has to be controlled in order to avoid the formation of an insulating layer between the electrode and the molecular catalyst. Several different linkers can be grafted on the same solid support and/or linked to identical metal-organic complexes. One linker arm can also be anchored to the solid support by two distinct functional groups. One linker arm can also bear more than one metal-organic complex, and those metal-organic complexes can be identical or different. Such variations are illustrated on the scheme B and C here-under:

The metal-organic complex which can be used in the material of the invention is selected from those responding to the following criteria:

The metal-organic complex is a molecule comprising one or several, preferably 1, 2 or 3 metal atoms and a ligand. The metal atoms are selected from the transition metals, and especially atoms 21 to 29, 42, 44 to 46 and 74 to 78 of the periodic classification of elements and an organic ligand. Those favourite transition metal atoms are: Sc, Ti, V, Cr, Mn, Fe, Co, Ni, Cu, Mo, Ru, Rh, Pd, W, Re, Os, Ir, Pt. The ligand can be constituted of one or several distinct molecules. The metal-organic complex, when studied in solution, before grafting on the linker, displays electrocatalytic activity for hydrogen evolution or uptake. This can be assessed simply by recording cyclic voltamograms in the solvent of interest in the presence of increasing amounts of a proton source (an acid with a strength comparable to that used in the aimed application) and observing current enhancements either cathodic for hydrogen evolution or anodic for hydrogen uptake, for overpotentials (related to the equilibrium potential of the H⁺/H₂ couple under the operating conditions) lower than 800 mV.

Preferably, electrocatalysis should rely on a reversible, pseudo-reversible or at least chemically reversible electrochemical process for the catalysts in an active state along the catalytic cycle. The metal-organic complex can be mono or polynuclear preferably with chelating, multidentate ligands for insuring stability of the coordination sphere. It is preferable for the coordination sphere of the metal not to contain a labile ligand or a ligand required for efficient catalytic properties that can be easily removed upon reduction or oxidation.

The metal-organic complex which can be used in the material of the invention could be selected from those, not limitating, responding to the following formulaes. More examples can be found in Coord. Chem. Rev. 2005 249, or Comptes Rendus Chimie 2008, 11, 8, two special issues devoted to hydrogenases and model compounds.

• Metal dioxime /diimine complexes such as those having the following formulaes:
• Metal amine /imine/pyridine complexes such as those having the following formulaes:
• Metal porphyrins and related (extended or restricted) polypyrrole macrocycles or metal phthalocyanine complexes or metal cyclam and related polyazamacrocyclic complexes such as those having the following formulaes:
• Ferrocenophane and related complexes such as those having the following formulaes
• Metal diphosphine or diphosphite complexes such as those having the following formulaes: wherein M represents an atom selected from transition metals of the periodic table of elements, and preferably, from the list which has been given above, and even more preferably M is selected from: Mn, Fe, Co, Ni, W, Mo, and advantageously M is selected from: Ni and Co.
  R₀, identical or different, represents a group selected from H, BF₂, BPh₂, BR¹₂, B(OR¹)₂,BFR¹.
  L is selected from a solvent molecule such as water, acetone methanol, ethanol, n-propanol, isopropanol, n-butanol, acetonitrile, tetrahydrofurane, dimethylformamide, dimethylsulfoxide, pyridine, an anionic ligand such as a halogen, a pseudohalogen such as SCN- or cyanide, hydride, oxygenated anion such as nitrate, sulphate, sulfonate, perchlorate, or a common monodentate ligand such as pyridine, imidazole, triazole, CO, H₂, formyl, phosphine, nitrile or isonitrile ligands, such as acetonitrile, benzonitrile, trimethyl isonitrile, benzylisonitrile.
  n is a number comprised between 0 and 6.
  R₁, R₂, R₃, identical or different, represent a group selected from : H, C₁-C₂₀ alkyl, C₁-C₂₀ alkenyl, C₁-C₂₀ alkynyl, C₆-C₃₀ aryl, C₆-C₃₀ aralkyl, possibly comprising one or more functions selected from: -OH, -NH₂, -COOH, -CONH₂, a triazole ring, possibly comprising one or more bridges selected from: -CO-O-CO-, -CO-NH-CO- and at least one of the R₁ groups, one of the R₂ groups or the R₃ group represents a covalent link with the linker molecule, two or more R₁ substituents can be fused together, two or more R₂ substituents can be fused together.

Preferably all the R₁ groups are identical with the possible exception of at least one R₁ group which is a covalent link with the linker molecule.

Preferably all the R₂ are identical with the possible exception of at least one R₂ group which is a covalent link with the linker molecule.

Preferably, the metal-organic complex is selected from metal diphosphine or diphosphite complexes such as those having the following formulaes, wherein R₁, R₂ and R₃ have the same definition as above:

A favourite variant is with the metal-organic complex selected from those which respond to the following formula (II):

Wherein R₁ and R₂ have the same definition as above.

Advantageously all the R¹ are identical and represent a group selected from: with the exception of at least one R₁ group which is a covalent link with the linker molecule.

Preferably, all the R₂ are identical and represent a phenyl group.

The functionalization of the solid support by the metal organic complex through linker arms can be done by several ways:

• (i) the metal-organic complex can be first coupled with the linker and this assembly can be grafted on the conductive or semi-conductive material with a high specific surface;
• (ii) the conductive or semi-conductive material with a high specific surface can be first functionalized with the linker and the metal-organic complex can be coupled afterwards on the linker using classical organic coupling reactions including methods of click chemistry;
• (iii) the conductive or semi-conductive material with a high specific surface can be first functionalized with a linker already coupled with the ligand which is part of the metal-organic complex. Then the complex is formed by reacting this grafted ligand with a mixture of metal salts or metallic precursors of the complex and free ligands.

Some of these methods are illustrated in the experimental part.

When the linker is the interaction between the linker and the solid support is via π-stacking. The linker is put in contact with the solid support and an interaction between the aryl group of the linker and the solid support is established. The linker may have been grafted or not, prior to this step, by the metal-organic complex C*, or by the ligand forming part of the metal-organic complex.

One example of the preparation of a material according to the invention, in the case wherein the linker is an oligomer resulting from the oligomerization of (IC), is illustrated on figure 14:

• A solid support, like a carbon nanotube support (CNT), is functionalized in a first step with a linker by oligomerization of

Another schematic representation of the preparation of a material according to the invention is illustrated on figure 15:

• A solid support is functionalized in a first step with a linker by oligomerization of

Another schematic representation of the preparation of a material according to the invention is illustrated on figure 16:

• In a first step, a linker precursor is functionalized by grafting of a metal-organic complex C* through reaction with the derivative C*W, wherein Ar, R, R', X, Y and x have the same meaning as above and W and Z represent reactive groups capable of forming together a covalent link.

A solid support is then functionalized with the linker-C* molecule resulting in oligomerization of the linker and establishment of a covalent link between the solid support and the oligomer.

The degree of oligomerisation is arbitrarily selected on these figures for the purpose of illustration and cannot be considered as a limitation.

The same variants can be applied to the case wherein the linker is with the difference that reaction between the linker precursor and the solid support will generally lead to a simple covalent bond and not the formation of a covalently-linked oligomer.

One example of an electrode based on a material of the invention is depicted on figure 13 and is a favourite object of the invention.

The assembly of the constituents of the electrode can be done using several pathways:

The conductive or semi-conductive material with a high specific surface can be first functionalized and then deposited on the conductive electrode material or, in a different order; deposition can be done first, followed by functionalization steps.

The material of the invention once deposited on a conductive support can be used as an electrode for electronic applications, like an electrode in fuel cells, especially hydrogen fuel cells or in an electrolyser, especially a water electrolyser. These new electrode materials show for the first time high rates of reduction of protons to hydrogen. Furthermore, the catalytic process occurs at low overpotentials. These materials combine the great surface area and conductivity of MWCNTs and the efficiency of molecular catalysts in reducing protons to hydrogen. Furthermore the favourite metal-organic complexes, nickel (II) diphosphine complexes of formula (II), are easy to synthesize, display great stability to air and are not poisoned by carbon monoxide.

The favourite surface modification technique is based on the reduction of functionalized aryldiazonium salts. This technique offers a versatile and controlled electrografting method for immobilization of redox catalysts onto electrodes. The high rate of heterogeneous electron transfer at the ITO/MWCNT or graphite/MWCNT electrode is maintained. Transparent electrodes based on CNTs can be obtained for applications in photoelectrochemical cells. (Trancik et al. Nano Letters 2008, 8, 982-987). This route towards modified electron is versatile and can be transposed to other high surface area carbon materials or metal oxide photoelectrodes. These low-cost and transparent electrodes can be readily integrated in energy conversion devices.

These air-stable nanostructured materials have proven their efficiency in the production of H₂ from H⁺ and could at least compete with platinum in electrocatalytic devices. The material can be used either in organic medium or in water. These air-stable nanostructured materials have proven their efficiency in the catalytic oxidation of H₂.

The new catalytic materials of the invention show numerous advantages:

• They use only a handful of metal atoms, by contrast to the particle approach. This can be done through the use of cheap, easy to synthesize ligands coordinating the metal atoms.

The catalytic activity of each metal atom is higher, so that the amount of metallic charge can be reduced.

Molecular catalysts are very selective towards their chemical substrate, as opposed to platinum nanoparticles that are irreversibly polluted by gas or organic pollutants.

The metals used are not necessarily noble metals.

The stability of the molecular catalyst is increased upon grafting by comparison with that reported in solution.

These new materials can be used as a cathodic active layer in a water electrolyser. As the metal-organic complex is known for being an active catalyst for hydrogen oxidation, these materials can also be used as anodic active layer in a hydrogen fuel-cell. More generally, this class of material could be used in any application requiring electrocatalysis provided the nature of the grafted catalyst has been adapted to the application (electroreduction of organic substrates such as ketone, oxygen formation from water, oxygen reduction for the cathodic active layer of a fuel cell, depollution of water using dechlorination of pollutants, CO₂ reduction....).

• Figure 1 illustrates the synthesis methods for the molecules of formula (I)
  • Scheme 1: Synthesis of P₂^PhN₂^R ( R = CH2C16H9, R = PhCH2CH2COOH)
  • Scheme 2: Synthesis of [Ni(P₂^PhN₂^R)₂](BF₄)₂
• Figure 2: ¹H and ³¹P NMR spectrum of [Ni(P₂^PhN₂^{PhCH2CH2COOPht})₂](BF₄)₂ in CD₃CN
• Figure 3: ¹H and ³¹P NMR spectrum of [Ni(P₂^PhN₂^CH2C16H9)₂](BF₄)₂ in CD₃CN
• Figure 4: Cyclic voltammograms of [Ni(P₂^PhN₂^R)₂](BF₄)₂ (1 mmol.L^-1, glassy carbon working electrode, 50mV.s^-1, MeCN-Bu₄NPF₆ 0.2mol.L^-1, potentials are quoted versus the Ag/AgClO₄, 0.01 mol.L^-1 electrode): full line : R = PhCH2CH2COOPht, dashed line : R = CH2C16H9 R, dotted line : R = Ph.
• Figure 5a (left): Cyclic voltammograms of [Ni(P₂^PhN₂^{PhCH2CH2COOPht})₂](BF₄)₂ (1 mmol.L^-1) in CH₃CN in the presence of 1.0, 1.5, 3.0, 5.0 and 10 equiv. of [DMFH]OTf (glassy carbon working electrode, 50mV.s^-1, MeCN-Bu₄NPF₆ 0.2mol.L^-1, potentials are quoted versus the Ag/AgClO₄, 0.01 mol.L^-1 electrode)
• Figure 5b (right): Cyclic voltammetry of a solution of [Ni(P₂^PhN₂^CH2C16H9)₂](BF₄)₂ (1 mmol.L^-1) in CH₃CN in the presence of 1.0, 1.5, 3.0, 5.0 and 10 equiv. [DMFH]OTf (glassy carbon working electrode, 50mV.s^-1, MeCN-Bu₄NPF₆ 0.2mol.L^-1, potentials are quoted versus the Ag/AgClO₄, 0.01 mol.L^-1 electrode)
• Figure 6: Trace of the ratio [I_p^c(H⁺)/I_p^c(Ni)] (where I_p^c(H⁺) is the peak current of the catalytic wave and I_p^c(Ni) is the peak current of a monoelectronic Ni-centered wave) towards the equivalents of [DMFH](OTf) for (▲) [Ni(P₂^PhN₂^{PhCH2CH2COOPht})₂](BF₄)₂ and (■) [Ni(P₂^PhN₂^CH2C16H9)₂](BF₄)₂
• Figure 7: XPS spectrum of ITO/MWNT-PhCH₂CH₂-NH₃⁺/-NH₂ electrode at N 1s binding energy level
• Figure 8: Cyclic voltammetry (3 successive scans) at an functionnalized ITO/MWCNT-PhCH₂CH₂NH-[Ni(P₂^PhN₂^PhCH2CH2CO)₂](BF₄)₂ electrode in MeCN-Bu₄NPF₆ 0.2 mol.L^-1 (50mV.s^-1; potentials are quoted versus the Ag/AgClO₄, 0.01 mol.L^-1 electrode)
• Figure 9 XPS spectrum of a functionalized ITO/MWCNT-PhCH₂CH₂NH-[Ni(P₂^PhN₂^PhCH2CH2CO)₂](BF₄)₂ electrode (Left : Ni 2p ; right : P 2p).
• Figure 10: Electrocatalytic behaviour of a (A) ITO/MWCNT electrode and (B) functionalized ITO/MWCNT-PhCH₂CH₂NH-[Ni(P₂^PhN₂^PhCH2CH2CO)₂](BF₄)₂ electrode in the presence of various concentrations of [DMFH]OTf (0, 0.16, 0.32, 0.50, 0.66, 1.0, 1.3, 2.0, 3.3, 6.6 mM) in MeCN-Bu₄NPF₆ 0.2M (50mV.s^-1; potentials are quoted versus the Ag/AgClO₄, 0.01 mol.L^-1 electrode)
• Figure 11: Trace of the ratio [I_p^c(DMFH⁺)/I_p^c(Ni)] towards the concentration of [DMFH](OTf) for a functionalized ITO/MWCNT-PhCH₂CH₂NH-[Ni(P₂^PhN₂^PhCH2CH2CO)₂](BF₄)₂ electrode.
• Figure 12: Coulometry at (A) an ITO/MWCNT electrode and (B) an ITO / MWNT-PhCH₂CH₂NH-[N₁(P₂^PhN₂^PhCH2CH2CO)₂](BF₄)₂ (area : 1cm^-1) poised at - 0.4 V vs Ag/AgClO₄ 0.01 mol.L^-1 in CH₃CN solution of [DMFH](OTf) (0.06 mol.L^-1) and n-Bu₄NBF₄ (0.1 mol.L^-1). The volume of the evolved hydrogen as a function of the charge is represented in the inset.
• Figure 13. Representation of the structure of an ITO/ MWNT-PhCH₂CH2NH-[Ni(P₂^PhN₂^PhCH2CH2CO)₂](BF₄)₂ electrode.
• Figure 14 to 16 schematically illustrate the steps fro the preparation of materials of the invention.
• Figurel7A and 17B: MEB image of a a functionalized ITO/MWCNT-PhCH₂CH₂NH-[Ni(P₂^PhN₂^PhCH2CH2CO)₂](BF₄)₂ electrode
• Figure 18: Result of cyclic voltametry experiments on materials of the invention

Materials. All reactions were routinely performed under an inert atmosphere of argon in a glove box or using standard Schlenk techniques. Solvents were degassed and distilled under argon. Diethyl ether was distilled by refluxing over Na/benzophenone; ethanol was distilled by refluxing over magnesium diethoxide; dry acetonitrile and dichloromethane were obtained by distillation on CaH₂. NMR solvents (Eurisotop) were deoxygenated by three freeze-pump-thaw cycles and stored over molecular sieves. Commercial dimethylformamide for electrochemistry was degassed by bubbling nitrogen through it. The supporting electrolyte (n-Bu₄N)BF₄ was prepared from (n-Bu₄N)HSO₄ (Aldrich) and NaBF₄ (Aldrich) and dried overnight at 80 °C under vacuum. [Ni(MeCN)₆](BF₄)₂ was prepared according to previously described procedures [B. J. Hathaway, A. E. Underhill, J. Chem. Soc. 1960, 3705; B. J. Hathaway, D. G. Holah, A. E. Underhill, J. Chem. Soc. 1962, 2444]. Other chemicals were used as received. Commercial 3100 grade Multi-walled Carbon Nanotubes (> 95 %) were obtained from Nanocyl. Carbon nanotubes were used as received, without any purification step. The ITO substrates were provided by Präzision Glas & Optik, GmbH (PGO).

NMR spectra were recorded at room temperature in 5 mm tubes on a Bruker AC 300 spectrometer equipped with a QNP probehead, operating at 300.0 MHz for ¹H, 75.5 MHz for ¹³C and 121.5 MHz for ³¹P or a Bruker AVANCE DRX 400 operating at 400.0 MHz for ¹H. Solvent peaks are used as internal references relative to Me₄Si for ¹H and ¹³C chemical shifts; ³¹P NMR spectra were proton-decoupled and referenced to H₃PO₄ (85%). ESI mass spectra were recorded with a Finnigan LCQ thermoquest ion-trap. Elemental analyses were done at the Microanalyses Laboratory at ICSN/CNRS, Gif/Yvette, France.

For Infrared spectroscopy (IR) a Bruker Vertex 70 spectrometer (resolution 2 cm^-1, spectra were collected with 256 scans, MCT detector), equipped with a Pike Miracle plate for ATR. UV-Vis spectra were recorded with a Perkin Elmer Lambda 650 spectrometer.

Electrochemical analysis was done using an EG&G potentiostat, model 273A. The electrochemical experiments were carried out in a three-electrode electrochemical cell under a highly controlled argon dry atmosphere in a glove box. The working electrodes were those described below. The auxiliary electrode was a platinum or graphite foil. The reference electrode was based on the Ag/AgClO₄ 10^-2 M couple which potential is +0.3V with respect to SCE. All potentials given in this work are with respect to the (Ag/AgClO₄) reference electrode.

The experiments were conducted in anhydrous dichloromethane or acetonitrile (water content <50 ppm) with tetrabutylammonium hexafluorophosphate [Bu₄N]PF₆ as the supporting electrolyte. Additions of [DMFH](OTf) were made by syringe as a 1:1 fresh mixture of dimethylformamide and triflic acid.

For bulk electrolysis experiments, a tight cell was used, connected to a U tube allowing volumetric monitoring of the evolved gas at atmospheric pressure. The platinum-grid counter electrode was placed in a separate compartment connected by a glass-frit and filled with the electrolytic solution. Hydrogen accumulated in the cell was tested for purity using a Delsi Nermag DN200 GC chromatograph equipped with a 3 m Porapack column and a thermal conductivity detector (TCD). Nitrogen under 1 bar was used as the carrier gas. The whole apparatus was thermostated at 45°C. Under these conditions, pure hydrogen has an elution time of 78 s.

X-ray photoelectron spectra (XPS) were recorded on a Kratos Analytical Axis Ultra DLD, using an Al Ka source monochromatized at 1486.6 eV. We used a hemispheric analyzer working at pass energy of 50 eV for the global spectrum, and 20 eV when focusing on the sole core levels.

All reagents and chemicals were purchased from Aldrich. Reagents and chemicals were used as received except when mentioned. [Ni(MeCN)₆](BF₄)₂ was prepared according to known procedures [B. J. Hathaway, A. E. Underhill, J. Chem. Soc. 1960, 3705; B. J. Hathaway, D. G. Holah, A. E. Underhill, J. Chem. Soc. 1962, 2444].

A 50% solution of tetrafluoroboric acid in Et₂O (0.5mL, 3.7 mmol) was added to a solution of 4-(2-aminoethyl)aniline (0.5mg, 3.7 mmol) in MeCN (4mL) at - 40°C under argon. The mixture was stirred 10 min, then nitrosyl tetrafluoroborate (450 mg, 3.8 mmol) was added dropwise. After 30 min of stirring at -40 °C Et₂O (100 mL) was added. The white precipitate obtained was filtered and washed twice with Et₂O to give a white powder (1.18 g, 99%).

¹H (CD₃CN): 3.26 (m, 2H, CH₂CH₂NH₃⁺), 3.31 (m, 2H, CH₂CH₂NH₃⁺), 7.82 (d, 2H, J = 9Hz, C₆H₄), 8.44 (d, 2H, J = 9Hz, C₆H₄)

820,781 IR: ν (cm^-1) = 3270, 3201, 3113, 2285, 1589, 1508, 1019, 928, 851, 820, 781

A solution of phenylphosphine 10% in hexane (3.3 mL, 3.0 mmol) was added dropwise to a solution of paraformaldehyde (180mg, 6.0 mmol, 2 equiv.) in degassed EtOH (40 mL). The mixture was heated under stirring at 80°C for 40 min. 3-(4-aminophenyl)propanoic acid (496 mg, 3.0 mmol, 1 equiv.) was added and the solution refluxed overnight. The solution was allowed to cool down to RT. A white precipitate (1.43 g, 2.4 mmol, 80%) was filtrated and washed with EtOH.

¹H (DMSO-d₆): 2.42 (t, 2H, CH₂CH₂COO, J = 7Hz), 2.67 (t, 2H, CH₂CH₂COO, J = 7Hz), 4.12 (m, 4H, NCH₂P), 4.54 (m, 4H, NCH₂P), 6.57-7.03 (m, 8H, NC₆H₄), 7.47-7.69 (m, 10H, PC₆H₅), 12.00 (s, 1H, COOH)

IR: ν (cm^-1) = 2961, 2921, 1708, 1613, 1515, 1456, 1435, 1379, 1243, 1190,799,788,746

A solution of P₂^PhN₂^PhCH2CH2COOH (320mg, 0.53 mmol), 1-ethyl-3-(3-dimethylaminopropyl)-carbodiimide (EDC, 191 mg, 1.23 mmol, 2.3equiv.) and N-hydroxyphtalimide (200 mg, 1.23 mmol, 2.3 equiv.) in DMF (2 mL) was stirred during 4h. Water (80mL) was added yielding a white precipitate which was filtrated and washed with acetone to give a white powder (310 mg , 0.35 mmol, 65%).

³¹P (CD₃CN): -52.6 (br s)

¹H (CD₃CN): 2.83 (t, 2H, CH₂CH₂COO, J = 7Hz), 2.95 (t, 2H, CH₂CH₂COO, J = 7Hz), 4.12 (m, 4H, NCH₂P), 4.53 (m, 4H, NCH₂P), 6.59-7.11 (m, 8H, NC₆H₄), 7.46-7.67 (m, 10H, PC₆H₅), 7.90-7.97 (8H, H Pht)

A solution of [Ni(MeCN)₆](BF₄)₂ (50 mg, 0.104 mmol) was stirred at RT during 3h in the presence of P₂^PhN₂^{PhCH2CH2COOPht} (200mg, 0.225 mmol, 2.2equiv.) in MeCN (10 mL). The solution was concentrated under vaccum and Et₂O was added to precipitate a red powder. After filtration, the red product (90mg, 0.045 mmol, 40%) was washed twice with Et₂O and dried in vacuum.

³¹P (CD₃CN): 3.71 (br s)

¹H (CD₃CN): 3.02 (m, 16H, CH₂CH₂COO), 3.93 (d, 8H, J = 5.6Hz, NCH₂P), 4.25 (d, 8H, J = 5.6Hz, NCH₂P), 7.13-7.23 (m, 16H, NC₆H₄), 7.32-7.39 (m, 20H, PC₆H₅), 7.7-7.9 (16H, H Pht)

IR: ν (cm^-1) = 1743,1508,1187,1057, 879, 697

UV/Vis (CH₃CN) : λ = 290nm (ε = 24000), 480 nm (ε = 1500)

Pyrenemethylamine hydrochloride (535 mg, 2.0 mmol) was added to a solution of NaOH (120 mg, 3.0 mmol) in water (40 mL). The resulting suspension was extracted with CH₂Cl₂ (3x20 mL). Removal of the solvent under vaccum yields a white power of pyrenemethylamine (415 mg, 1.80 mmol).

A solution of phenylphosphine 10% in hexane (1.9 mL, 1.72 mmol) was added dropwise to a solution of paraformaldehyde (100 mg, 3.33 mmol 2 equiv.) in EtOH (40mL),. The mixture was heated under stirring at 80°C for 40 min. Pyrenemethylamine (400 mg, 1.72 mmol, 1 equiv.) was added and the solution was refluxed overnight. Upon cooling to RT, a white precipitate (1.07 g, 1.46 mmol, 85%) formed which was filtrated and washed with EtOH.

³¹P (CD₃CN): -48.0 (br s)

IR: ν (cm^-1) = 3044, 2872, 2830, 1250, 1068, 837, 739, 711

A solution of [Ni(MeCN)₆](BF₄)₂ (70 mg, 0.14 mmol) was stirred at RT during 3h in the presence of P₂^PhN₂^{PhCH2CH2COOPht} (200mg, 0.2mmol, 2equiv.) in MeCN (10 mL). The solution was filtrated and concentrated under vaccum. Et₂O was added to precipitate an orange powder. After filtration, the red-orange product (220 mg, 0.13mmol, 90%) was washed twice with Et₂O and dried in vacuum.

³¹P (CD₃CN): 3.37 (br s)

¹H (CD₃CN): 3.14 (br d, 8H, J = 20Hz, NCH₂P), 3.42 (d, 8H, J = 12 Hz, NCH₂C₁₆H₉), 4.50 (br d, 8H, J = 55Hz, NCH₂P), 6.48-6.92 (m, 20H, PC₆H₅), 7.73-8.04 (36H, C₁₆H₉)

IR: ν (cm^-1) = 3061, 1592, 1494, 1194, 1050, 746, 690

UVNis (CH₃CN) : λ = 313 nm (ε = 51000), 328 nm (ε = 105000), 344 nm (ε = 130000), 471 nm (ε = 1400)

ITO/MWCNT electrodes were made by filtration technique on mixed cellulose ester membrane of a suspension of MWCNT (0.1 mg) in water (80 mL) [2]. The membrane was deposited on an ITO wafer and dissolved through washings with MeCN and acetone. Samples were further heated at 100°C.

Graphite/MWCNT electrodes were made by filtration technique on mixed cellulose ester membrane of a suspension of MWCNT (0.1 mg) in water (80 mL) (Wu et al., Science 305, 2004, 1273-1276). The membrane was deposited on a graphite wafer and dissolved through washings with MeCN and acetone. Samples were further heated at 100°C.

PTFE/MWCNT electrodes were made by filtration of a suspension of MWCNT in EtOH on a PTFE membrane yielding a so-called PTFE/MWCNT bucky paper.

The ITO/MWCNT electrodes were used as working electrodes in a three-electrode cell in the presence of 4-(2-Ammonioethyl)benzenediazonium tetrafluoroborate (1 mmol.L^-1). Electrografting was performed though reduction of the diazonium salt. This can be done either using controlled potential coulometry at E_p = - 0.4V or cyclic voltammetry (recording 3 cycles at 20mV.s^-1 between 0.4V and -0.4V).

The Ni complex was introduced by a post-functionalization step based on the formation of an amide linkage between the amine residue introduced at the surface of the MWCNT electrode and the activated ester group of [Ni(P₂^PhN₂^{PhCH2CH2COOPht})₂](BF₄)₂. This step was achieved by dipping the wafer in a solution of [Ni(P₂^PhN₂^{PhCH2CH2COOPht})₂](BF₄)₂ (0.5 mmol.L^-1) in CH₃CN in the presence of Et₃N (2 mmol.L^-1). The solution was stirred overnight. XPS and electrochemistry experiments confirmed the covalent grafting of Ni complexes to MWCNTs. The structure of the grafted electrode can be represented by the scheme illustrated on figure 13.

Functionalized PTFE/MWCNT bucky papers were directly prepared by filtration of a suspension of MWCNTs (0.1 mg) in a EtOH/CH2Cl2 mixture (80/2 v:v, 80 mL) in the presence of [Ni(P2PhN2CH2C16H9)2](BF4)2 (5 mg). Π-stacking of the Ni complex bearing pyrene groups was conformed by XPS analysis and electrochemistry.

An activated ester P₂^PhN₂^R (R3 = PhCH₂CH₂COOPht) was synthesized by coupling the carboxylic function in P₂^PhN₂^R1 with N-Hydroxy-phtalimide in the presence of EDC. The synthesis of the nickel complexes [Ni(P₂^PhN₂^R)₂](BF₄)₂ (R = R2 and R3) was achieved by mixing 2 equivalents of the diphosphine ligand and a nickel(II) precursor in MeCN (Figure 1 - scheme 2). The ligands and the complexes were characterized by ³¹P NMR and ¹H NMR. The chemical shifts of both complexes are similar to those obtained by Dubois for [Ni(P₂^PhN₂^Ph)₂](BF₄)₂ (Figures 2 and 3).

Two reversible reductions at E_1/2^red1 = -0.75 - - 0.8 and E_1/2^red2 = - 0.95 - -1.05 V vs Ag/AgClO₄, 0.01 mol.L^-1 (Table 1) were observed in MeCN at a glassy carbon electrode (Figure 4). The first reduction exhibits a slow electron transfer compared to the second one. This is probably due to a ligand reorganization triggered by the first electron transfer slowing down the heterogeneous transfer rate. [Ni(P₂^PhN₂R)₂](BF₄)₂ E_1/2^red1 (V) E_1/2^red2 (V) R=Ph - 0.77 - 0.95 R = PhCH2CH2COOPht - 0.76 - 0.95 R = CH2C 16H9 - 0.80 - 1.05

Addition of acid to the electrochemical cell confirmed their electrocatalytic behavior towards the reduction of proton. At E_p^red - -0.5 V vs Ag/AgClO₄, 0.01 mol.L^-1, a reduction wave is detected, indicating the protonation of the Ni(II) complex. Electrocatalytic reduction was confirmed by cyclic voltammetry performed at different concentrations of acid [DMFH]OTf (Figure 5). The linear dependence of the catalytic current with the number of added protons (Figure 6) confirmed the electrocatalytic nature of the reduction process and showed that the modifications of the ligands did not affect the properties of the catalysts.

The ITO/MWCNT and graphite / MWCNT electrodes were made by deposition of a thin film of nanotubes on ITO or graphite with a soluble membrane technique. (Z. C. Wu et al., Science 2004, 305, 1273). These electrodes were modified by reduction of a diazonium salt (4-(2-Ammonioethyl)benzenediazonium tetrafluoroborate) at the ITO/MWCNT or graphite / MWCNT working electrode. Cyclic voltammetry was used to control the reduction of the diazonium salt. An irreversible reduction is observed at E_p^red = 0.26 V vs Ag/AgClO₄, 0.01 mol.L^-1, corresponding to the reduction of 2-aminoethylbenzenediazonium and the grafting of the polyphenylene layer on the CNT electrode. Several scans were performed to warrant a good functionalization of the CNTs. On the second and third scan, a deviation of the reduction peak to more negative potentials indicate an increasing thickness of the polyphenylene layer but a conserved fast electron transfer. Different conditions were studied by controlled potential electrolysis to maximize the quantity of amine groups at the electrode surface. The reversible reduction system of the immobilized nickel complex (see below) was used as a redox probe to determine the best conditions for obtaining high surface concentrations of the catalyst.

The functionalized CNT surface was further characterized by XPS measurement. XPS spectrum of the energy of 1s orbitals of N atom exhibits two peaks at 399.9 and 401.6 eV, indicating the presence of both amine and ammonium groups. (Figure 7)

An heterogeneous amidification reaction was realized between functionalized electrodes and nickel complexes [Ni(P₂^PhN₂^R3)₂](BF₄)₂ bearing an activated ester group. The amine-functionalized electrodes were dipped in a solution of [Ni(P₂^PhN₂^R)₂](BF₄)₂ in the presence of triethylamine. The solution was stirred overnight. The covalent grafting of the complex was confirmed by cyclic voltammetry (Figure 8) and XPS measurements (figure 9). The structure of the final electrode material is represented in scheme 3.

The XPS spectrum (figure 9) exhibits a clear Ni 2p_1/2 and 2p_3/2 signal corresponding to nickel atoms in a high oxidation state.

Cyclic voltammetry (Figure 8) confirms the presence of the nickel covalently grafted to the electrode. A reduction at E_1/2^red- -0.95V (ΔE = 140mV) is observed, corresponding to the second reduction of the complex in solution. Resistivity due to the polyphenylene layer prevents from distinguishing the first slow electron transfer and the first reduction is not clearly detected. Considering the planar area of the electrode, the surface concentration of the nickel complex was estimated from the charge quantity calculated from the integration of the oxidation or reduction wave. Average concentrations were of an order of 10^-9 mol.cm^-2. The high surface area of the electrode was confirmed by the MEB image of these nanostructured electrodes (figure 17A and 17B).

When [DMFH](OTf) was added as the proton source, a reduction peak appeared at a potential that is 300mV more negative as compared to that achieved in the absence of acid and corresponding to the reduction of the protonated immobilized Ni(II) complex.

When the concentration of acid was increased, the peak current increased proportionally (Figure 10). A supported electrocatalytic behaviour is doubtlessly evidenced by the trace of the catalytic current at different concentrations of acid (Figure 11).

Electrolysis experiments were performed to investigate the stability of the covalent linkage and the effective production of H₂. Electrolysis was performed at a potential of -0.5V vs Ag/AgCl in an electrochemical cell coupled to a GC chromatographer. The catalyst activity is stable for hours. A charge current decrease is observed on a long time experiment for ITO/MWCNT electrodes that is due to the degradation of the CNT-ITO interface. This degradation does not happen for graphite/MWCNT based electrodes. 0.3mL of H₂ were produced and characterized by GC in 1h (Figure 12, see the experimental section). Taking into account a blank experiment in order to subtract residual current passing through the cathode under the same conditions but in the absence of the grafted nickel catalyst this correspond to the passage of 3 C. Considering the surface concentration of the catalysts, turnover numbers were estimated to be 15000 within a 1 hour experiment corresponding to turnover frequencies of 6 s^-1. The production of H₂ thus proved to be highly efficient. By comparison, the nickel complexes in solution were only shown to achieve 7 turnovers under similar bulk electrolysis conditions (DuBois et al., J. Am. Chem. Soc. 2006, 128, 358)

It is noteworthy that catalytic behaviour for pristine and amino-functionalized CNTs was observed at -1.2V, i.e. 500mV more negative than the nickel-functionalized-MWCNT. This study implicates the great importance of the presence of catalyst at the surface of the electrode so that hydrogen can be produced at low overpotentials.

Electrocatalysis is also observed for the same material when assayed in 0.5 mol.L^-1 H₂SO₄ aqueous solution. Cyclic voltametry experiments (figure 18) showed cathodic current corresponding to hydrogen evolution at -0.1V vs NHE, while the same unfunctionalized electrode material requires a potential 200 mV more negative to reach the same current intensity